Visual Analytics of Dynamic Networks

Total Page:16

File Type:pdf, Size:1020Kb

Visual Analytics of Dynamic Networks Visual Analytics of Dynamic Networks Paolo Federico Dissertation Faculty of Informatics, TU Wien, May 31, 2017 Visual Analytics of Dynamic Networks DISSERTATION submitted in partial fulfillment of the requirements for the degree of Doktor/in der technischen Wissenschaften by Paolo Federico Registration Number 0928613 to the Faculty of Informatics at the Vienna University of Technology Advisor: Univ.Prof.in Dr.in rer.soc.oec. Silvia Miksch The dissertation has been reviewed by: (Ao.Univ.Prof.in Dr.in phil. (Univ.Prof. Dr.rer.nat. Margit Pohl) Dr.techn.h.c. Dr.-Ing.E.h. Thomas Ertl) Wien, 25.04.2017 (Paolo Federico) Technische Universität Wien A-1040 Wien ⇧ Karlsplatz 13 ⇧ Tel. +43-1-58801-0 ⇧ www.tuwien.ac.at Acknowledgements I would like to thank my advisor Silvia Miksch for supporting me during the course of my doc- toral study with scientific rigour, friendly communication, and relentless enthusiasm into visual analytic research. Thanks to her and to all my present and past colleagues in the ieg/CVAST group, I could work on my research in an enjoyable, inspiring, and supportive work environ- ment. In addition, I would like to thank the researchers who collaborated to the projects in whose context I conducted my PhD, namely Wolfgang Aigner, Albert Amor-Amorós, Jürgen Pfeffer, Michael Smuc, Florian Windhager, and Lukas Zenk. Special thanks go to Wolfgang Aigner who provided fruitful suggestions and feedback in the first phases of my PhD. Christian Bors and Markus Bögl helped me with the German abstract. The research leading to this thesis has received funding from the Austrian Research Pro- motion Agency (FFG) through ViENA (Visual Enteprise Network Analytics, project number 820928) and Expand (EXploratory visualization of PAtent Network Dynamics, project number 835937) as well as from the Austrian Federal Ministry of Science, Research, and Economy (for- merly known as Austrian Federal Ministry of Economy, Family and Youth) through the Laura Bassi Centre for Visual Analytics Science and Technology (CVAST), project number: 822746 (Phase 1). iii Abstract While there are many well-established techniques to analyze and visualize static social networks, visual analysis of dynamic (i.e., time-oriented) network data emerged in recent years as a rel- evant research topic, facing several open problems. The dynamic nature of this kind of data, indeed, poses the challenge of understanding both its relational aspect (the structure of social interactions) and its temporal aspect (how they change over time). In this doctoral work, we investigate how a visual analytics approach, integrating automatic analysis, visualization, and user interaction techniques, can support the examination of such dy- namic networks. In particular, by focusing on this research problem, we present the following contributions: 1. we propose a set of novel metrics (change centrality metrics) to specifically analyse how the network structure changes over time; 2. we combine different visual encodings for the time-oriented aspect of network data, enabling smooth transformations between differ- ent views; 3. we introduce novel techniques for user interaction, such as interactive control of dynamic layout stability and the vertigo zoom, allowing seamless transitions between relational and temporal perspectives on dynamic network data. We illustrate our approach by describing a prototypical implementation and demonstrate its utility by introducing a real-world usage scenario. Furthermore, we provide a validation of our approach by reporting findings from expert reviews (involving experts from both the visualization community and the problem domain) as well as from two task-based user-studies, namely a qualitative evaluation and a quantitative controlled experiment. These findings afford an indication of the overall validity of our approach and allow us to discuss how particular techniques and their combinations can support specific analytical tasks on dynamic network data. v Contents I The problem 1 1 Introduction 3 1.1 Motivation . 3 1.2 Research questions . 4 1.3 The visual analytics approach . 4 1.4 Research methodology . 6 1.5 Structure of this document . 12 2 Foundations and state of the art 13 2.1 Concepts and definitions . 13 2.2 Data models . 18 2.3 Automated analysis . 22 2.4 Historical graph visualization . 25 2.5 Graph visualization: surveys, taxonomies, design spaces . 29 2.6 Visual encodings for graphs . 32 2.7 Visual encodings for temporal graphs . 50 2.8 Interaction . 65 2.9 Task taxonomies . 68 2.10 Evaluation of graph visualization . 69 2.11 Limitations of existing approaches and open challanges . 71 II The proposed solution 73 3 Analysis 75 3.1 Automated analysis of static networks . 75 3.2 Automates analysis of dynamic networks . 77 4 Visualization 83 4.1 Visual encoding . 83 4.2 Dynamic layout . 83 4.3 Views . 84 4.4 Enriching visualization with analysis results . 87 ix 4.5 Exploiting change centrality metrics . 88 5 Interaction 93 5.1 Basic interactions . 93 5.2 Smooth animated transitions between views . 94 5.3 Interactive control of layout stability . 94 5.4 Dual-mode highlighting . 95 5.5 Trajectories on demand . 97 5.6 Switching between relational and temporal perspectives . 99 6 Implementation notes 105 III The validation 107 7 Usage Scenario 109 7.1 Analysis of network structure . 110 7.2 Hires, leaves, and resignations . 111 7.3 The trend of individual performance . 112 7.4 Presence of key players and their evolution . 112 8 Expert Review 115 9 Qualitative User Study 117 9.1 Usability findings . 118 9.2 Task Completion Analysis . 119 9.3 Multiple Problem Solving Strategies . 120 10 Quantitative User Study 125 10.1 Study design . 125 10.2 Stimuli . 127 10.3 Tasks . 127 10.4 Subjects’ pool and study settings . 128 10.5 Hypotheses . 128 10.6 Analysis . 129 10.7 Results . 129 IV Conclusion 133 11 Conclusion 135 11.1 Summary of contributions . 135 11.2 Answers to research questions . 137 11.3 Future directions . 138 11.4 Publications and dissemination . 139 x Bibliography 143 xi Part I The problem 1 CHAPTER 1 Introduction 1.1 Motivation Networks are exploited to model diverse phenomena in various domains: social interactions between human beings (sociology), as well as digital connections between electronic devices (communications), relationships between proteins (biology), and interdependencies of industrial sectors or regional markets (economics). Dynamic networks take into account changes over time: they not only model relations between different entities, but also consider the evolution of these relations, i.e. the way and the extent by which they change over time. Dynamic social networks, in particular, can be useful to model and analyze human rela- tionships in several potential scenarios: the informal social relationships of individuals within a family or a group of friends; the widespread connections through social networking services; the covert activities of small, interconnected terrorist cells; or the structured collaboration of employees within an enterprise. Visual analysis of dynamic social networks is a topic that has been drawing increasing atten- tion in recent years, not only from different research communities (not limited to social science researchers) but also from the general public. One reason for this interest is the availability of data: nowadays electronic devices often mediate interpersonal interactions or, because of their presence in our every day lives, are in any case able to capture and store social network data. Another reason is the possibility to gain objective insights about social relationships, in order to monitor and understand them, as well as take action to improve or better exploit them. Whatever application domain we consider, a good visualization of dynamic networks has to support the analysis of the relational aspect (what is the structure of the network) as well as the temporal aspect (how the network evolves over time), and should also enable a seamless switching between the two perspectives. In this work, we specifically consider the organizational network (i.e. a network consisting of the employees of an organization) of a knowledge intensive enterprise and focused on different kinds of relations, such as communication, collaboration, technical and practical advice, and spreading of new ideas. We aim to support users analyzing the evolution of these relations 3 as well as some performance indicators, by also considering how they relate to organizational changes: turnover, team restructuring, and other management actions. Analysis of social, i.e., interpersonal, networks is obviously a sub-field of sociology; it uti- lizes concepts from graph theory, combined with data models and computational algorithms, to perform automated data analysis. Graph drawing and information visualization provide algo- rithms and techniques for static and interactive visualization of network data. In this work, we tackle the problem from a multidisciplinary perspective, abstracting dy- namic network data from the specific problem domain (i.e., social networks), and investigating a visual analytics approach, as a combination of automated analysis and interactive visualization. 1.2 Research questions In particular, our research aims at investigating the following research question: How can a visual analytics approach support the examination of dynamic networks ac- • cording to specific user tasks? This main research question can be further detailed by three interconnected sub questions: – How can temporal aspects of network data
Recommended publications
  • Network Science
    This is a preprint of Katy Börner, Soma Sanyal and Alessandro Vespignani (2007) Network Science. In Blaise Cronin (Ed) Annual Review of Information Science & Technology, Volume 41. Medford, NJ: Information Today, Inc./American Society for Information Science and Technology, chapter 12, pp. 537-607. Network Science Katy Börner School of Library and Information Science, Indiana University, Bloomington, IN 47405, USA [email protected] Soma Sanyal School of Library and Information Science, Indiana University, Bloomington, IN 47405, USA [email protected] Alessandro Vespignani School of Informatics, Indiana University, Bloomington, IN 47406, USA [email protected] 1. Introduction.............................................................................................................................................2 2. Notions and Notations.............................................................................................................................4 2.1 Graphs and Subgraphs .........................................................................................................................5 2.2 Graph Connectivity..............................................................................................................................7 3. Network Sampling ..................................................................................................................................9 4. Network Measurements........................................................................................................................11
    [Show full text]
  • A Comparison of Self-Organizing Maps and Pathfinder Networks for the Mapping of Co-Cited Authors Jan William Buzydlowski Howard D
    A Comparison of Self-Organizing Maps and Pathfinder Networks for the Mapping of Co-Cited Authors A Thesis Submitted to the Faculty of Drexel University by Jan William Buzydlowski in partial fulfillment of the requirements for the degree of Doctor of Philosophy December 2002 © Copyright 2003 Jan William Buzydlowski. All Rights Reserved ii Dedications To my father. iii Acknowledgements Howard D. White, PhD, thesis advisor, who never treated me as a student, but from whom I learned more than any other teacher. The distinguished members of my committee: · Xia Lin, PhD, who treated me as a colleague and friend. · Katherine W. McCain, PhD, who always knew what to do. · Richard Heiberger, PhD, who showed me fascinating new worlds. · Stephen Kimbrough, PhD, for his invaluable contribution. J. Jeffery Hand, Linda Marion, Joan Lussky for their camaraderie. All of the experts involved with this study who were more than generous with their time and expertise. Cheryl L. Berringer for her emotional support, without which none of this would have been possible. iv Table of Contents List of Tables ......................................................................................................... vii List of Figures ........................................................................................................viii Abstract ....................................................................................................................x 1. INTRODUCTION AND STATEMENT OF PROBLEM ...................................1 1.1. Goal of the Research.............................................................................1
    [Show full text]
  • Cognitively Plausible Information Visualization
    Exploring Geovisualization J. Dykes, A.M. MacEachren, M.-J. Kraak (Editors) q 2005 Elsevier Ltd. All rights reserved. Chapter 35 Cognitively Plausible Information Visualization Sara Irina Fabrikant, Department of Geography, University of California Santa Barbara, Santa Barbara, CA 93106, USA Andre´ Skupin, Department of Geography, University of New Orleans, New Orleans, LA 70148, USA Keywords: Information Visualization, spatialization, Cartography, spatial cognition, human–computer interaction Abstract Information Visualization is concerned with the art and technology of designing and implementing highly interactive, computer supported tools for knowledge discovery in large non-spatial databases. Information Visualization displays, also known as information spaces or graphic spatializations, differ from ordinary data visualization and geovisualization in that they may be explored as if they represented spatial information. Information spaces are very often based on spatial metaphors such as location, distance, region, scale, etc., thus potentially affording spatial analysis techniques and geovisualization approaches for data exploration and knowledge discovery. Two major concerns in spatialization can be identified from a GIScience/ geovisualization perspective: the use of space as a data generalization strategy, and the use of spatial representations or maps to depict these data abstractions. A range of theoretical and technical research questions needs to be addressed to assure the construction of cognitively adequate spatializations. In the first part of this chapter we propose a framework for the construction of cognitively plausible semantic information spaces. This theoretical scaffold is based on geographic information theory and includes principles of ontological modeling such as semantic generalization (spatial primitives), geometric generalization (visual variables), association (source–target domain mapping through spatial metaphors), and aggregation (hierarchical organization).
    [Show full text]
  • FINDING PATHS OR GETTING LOST? – Examining the Mental Model Construct and Mental Model Methodology
    Linköping University | Department of Computer Science Master’s thesis, 30 credits | Cognitive Science Spring 2017 | ISRN LIU-IDA/KOGVET-A--17/009--SE FINDING PATHS OR GETTING LOST? – Examining the mental model construct and mental model methodology Jimmy Hammarbäck Supervisor: Rita Kovordányi External supervisor: Jonas Hallberg Examiner: Arne Jönsson Linköping University SE-581 83 Linköping, Sweden +46 013 28 10 00, www.liu.se Upphovsrätt Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art. Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart. För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/. Copyright The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non- commercial research and educational purpose.
    [Show full text]
  • Visualizing Knowledge Domains1
    This is a preprint of Katy Börner, Chaomei Chen, & Kevin Boyack: Visualizing Knowledge Domains. In Blaise Cronin (Ed.), Annual Review of Information Science & Technology, Volume 37, 1 Medford, NJ: Information Today, Inc./American Visualizing Knowledge Domains Society for Information Science and Technology, chapter 5, pp. 179-255, 2003. Katy Börner School of Library and Information Science, Indiana University, Bloomington, IN 47405, USA [email protected] Chaomei Chen College of Information Science and Technology, Drexel University, Philadelphia, PA 19104, USA [email protected] Kevin W. Boyack Sandia National Laboratories, Albuquerque, NM 87185, USA [email protected] "The purpose of computing is insight – not numbers." R. W. Hamming (1962) 1 INTRODUCTION ...............................................................................................................................................2 2 HISTORY.............................................................................................................................................................4 2.1 SCIENTOMETRICS, BIBLIOMETRICS, AND CITATION ANALYSIS ......................................................................4 2.2 MAP GENERATION AND VISUALIZATION ........................................................................................................7 3 PROCESS FLOW OF VISUALIZING KNOWLEDGE DOMAINS .............................................................8 3.1 UNITS OF ANALYSIS .....................................................................................................................................10
    [Show full text]
  • Selected Research and Development Papers Presented at the Annual
    thannual Proceedings Selected Research and Development Papers Presented at The Annual Convention of the Association for Educational Communications and Technology - Volume 1 Sponsored by the Research and Theory Division And The Division of Instructional Design 37 Jacksonville, FL Editor: Michael Simonson Nova Southeastern University, North Miami Beach, Florida 2014 Annual Proceedings - Jacksonville: Volumes 1 & 2 Volume 1: Selected Research and Development Papers And Volume 2: Selected Papers On the Practice of Educational Communications and Technology Presented at The Annual Convention of the Association for Educational Communications and Technology Sponsored by the Research and Theory Division And The Division of Instructional Design Jacksonville, FL 2014 Editor Michael Simonson Professor Instructional Technology and Distance Education Nova Southeastern University Fischler School of Education and Human Services North Miami Beach, FL Preface Since 1979, the Research and Division of the Association for Educational Communications and Technology (AECT) has sponsored or co-sponsored the publication of these Proceedings. Papers published in this year's 37th Annual Proceedings were presented at the 2014 AECT Convention in Jacksonville, FL.A limited quantity of these Proceedings were printed and sold in both hardcopy and electronic versions. Volumes 1 and 2 are available through the Educational Resources Clearinghouse (ERIC) System. Proceedings volumes are available to members at AECT.ORG. The Proceedings of AECT’s Convention are published in two volumes. Volume #1 contains papers dealing primarily with research and development topics. Papers dealing with the practice of instructional technology including instruction and training issues are contained in Volume #2. This year, both volumes are included in one document. REFEREEING PROCESS: Papers selected for presentation at the AECT Convention and included in these Proceedings were subjected to a reviewing process.
    [Show full text]